Google promoted Texas gunman fake tweets | Hacker News Reader

Google promoted Texas gunman fake tweets

bbc.co.uk

117 pointsdberhane137 comments

My summary of the article: An algorithm designed by Google engineers to promote upcoming stories does exactly what it was supposed to do, and people who don't understand why Google employees did not manually review just one out of one hundred thousand things that the search engine indexed that day are suprised.

Google Search is a content aggregator that show you what it thinks you are most likely to click on. It does not know about politics, it cannot fact check, it does not think care about effort journalism, the only thing that matters is what stories generate more ad revenue.

Of course, Google will try to police its results better after this incident, but they can't effectively do this without mass censorship. If you don't believe me, see the recent YouTube drama over monetization.

Maybe algorithms just aren't up to the task of exercising editorial judgement over news stories?

Would the world be that much poorer if Google didn't include a "popular on Twitter" feature?

What if Google News were a smaller, human-curated roster of important stories that were basically the same for everyone? (maybe targeted by country) It could be a good thing to have a non-filter bubbled set of news.

It's worth pointing out that Facebook tried the human curation thing and were accused of bias.

That's fine: They should own that, then. "We're Facebook and these are our values -- we trust the BBC and do not trust InfoWars.com"

These big tech companies are trying to be all things to all people, and they want to pretend that using algorithms means they're free from bias. It's just a different, more opaque bias encoded in the algorithms.

But users are inclined to trust it because it comes "from the computer". It's a Garbage In, Gospel Out situation.

You mean like every human-curated news source since the beginning of time?

Human curators are biased. Google company culture is left-leaning as a whole, so you'd expect to see that same bias in any curated content they provide.

>My summary of the article: An algorithm designed by Google engineers to promote upcoming stories does exactly what it was supposed to do, and people who don't understand why Google employees did not manually review just one out of one hundred thousand things that the search engine indexed that day are suprised.

That is not entirely correct, big dotcoms do outsource a lot of such mechanical Turk work to outsourcing companies that do such work, and they do prefer to contract ones in places without much contact to English speaking work to prevent mass backlash.

>https://www.wired.com/2014/10/content-moderation/

A lot of times when such companies claim that something is done by some voodoo magic "smart algorithm," that claim is a lie.

* s/work/world/

I read it more as Google's blind faith in algorithms finally being loudly and publicly discredited. Baked deeply into Google's corporate DNA is the unquestioned assumption that all you have to do to solve a problem -- any problem -- is throw clever algorithms and compute power at it. Which I suppose might still turn out to be true, but Google has picked off all the low-hanging fruit and is now hitting problems that seem to be well beyond their capabilities to solve algorithmically. But since the whole company is built on the assumption that this would never happen, the are incapable of confronting and dealing with it.

Algorithms can have bias, too. Like for instance, in the last election, Google removed the more negative expressions that would be suggested alongside politicians' names. That wasn't just the Algorithm™ -- it was Google deciding that.

There are likely hundreds of such "manual" decisions within its search algorithms.

Algorithms are generally incapable of bias.

What an algorithm does is a perfect reproduction of it's instructions plus it's parameters.

The actual bias is either by the designer by making an algorithm that does not fit the intended use properly or by using flawed parameters.

Also known as; garbage in, garbage out. The pipe in between doesn't care if it's garbage.

In your example, this was a decision by a Google engineer or c-level, not the algorithm. The algorithm did as it was told to in exactly the way it was designed without bias or error.

> Algorithms are generally incapable of bias.

This is nonsense unless you define-down "bias" in such a way as to mean "performing contrary to the deterministic result of its implementation" (which seems to be what you're doing, and I agree is impossible).

A more sensible definition of bias might be something like "preferentially favouring one outcome over another for reasons unrelated (or even contrary) to the core purpose of the software". Imagine a vending machine that when prompted generically by a user saying "I'm thirsty" always provided sugary soda instead of bottled water.

Modern machine learning algorithms are extremely complex and opaque, and depending on the dataset they're trained on, and the features they're trained to focus on, they can very easily wind up just encoding human biases.

https://www.theregreview.org/2017/10/10/cramer-bias-computer...

Yes - the fundamental limits to learning by induction

I don't think it's blind faith in algorithms, but that Google is using algorithms, because the only other alternative is to hire 100,000 people to do unrewarding grunt work classifying this stuff.

They are trying to provide more services using the tools they've develop. Maybe they've exceeded their reach, but I wouldn't blame that on a "blind faith" in algorithms.

And no, I am not a fan of Google. I think they are evil.

What you're saying is that some things are (presently, at least) impossible to automate reliably. That's true. The issue is that a reasonable person's response is "I guess we can't automate that yet," not "oh well, let's automate it really poorly instead."

You forget the part where the aggregated content is displayed in a different context than the one it is created for.

It's perceived as an factual answer to a question because it shows up when googled with keywords other than "Popular on Twitter" : https://twitter.com/justinhendrix/status/927335154707828736/...

When you see these tweets on Twitter, you'll quickly identify them as fake due to the context. On Google on the other hand you will see the false statements but these false statements would not be corrected if these tweets were not the exact content you're looking for and you just consumed them and moved on to the content you were searching.

So, a product is badly broken, and you're trying to sell this bug as a feature - not just of the product itself, but of the world we live in?

This is something that programmers (including myself) are often guilty of. We see an implementation and understand what features based on that implementation are difficult/infeasible. Then we say things like "it isn't possible with our architecture" and "It is understandable that the website works in $x way."

The flipside is true too: developers are very understanding of bugs/shortcomings that are borne out of the specific implementation. With all this knowledge, it is easy to say things like "people shouldn't use this tool for that" or "users should know this is flaky"

At some point you have to fight for the user instead of defend the code.

No, I suppose my point was more that people think that one of the features of Google search results is correctness, which is false. The main feature of Google searches is that it shows you what it thinks you want to see.

The "product" here is the fast search engine that indexes a massive portion of all human knowledge and tries to provide you with results that are meaningful to you all within a quarter of a second. I don't think it is broken, I think it is a miracle that it works as well as it does.

Unfortunately, there is no algorithm to determine "the truth", mainly because there is no universal truth. This is a feature of the world we live in, though I suppose you could call it a bug if you think that there should be some kind of universal truth.

If nearly all of your users believe that your product delivers X, when it actually delivers Y, then your product is still broken.

No, your product is not broken. You have a marketing problem. You sold your product as something it was not and now you pay the price of lying to all your users.

I bought a car, when I turn the steering wheel left, the wheels turn right. Damn marketing!!!

More like, "This piece of software I don't understand did something I was not expecting, so now I'm upset that the software isn't actually what I thought it was and I want it changed".

If your software is actively presenting something as X, when it's actually Y, that's not a marketing issue. That's your software displaying erroneous data. That's a bug.

There is explicit marketing and implied (historical, metaphorical, etc) behavior.

That Google doesn't intend to disseminate fake news does not exculpate them if that is nonetheless a foreseeable consequence of their algorithm.

How can you determine something IS a fake news? I can post an event I witnessed solely on my own, and yet people can call it fake news. Is UFO a fake news? No scientists have disapproved the possibility of the existence of UFO. I am sure many of the UFO accounts aren't UFO, glare or whatnot. But you cannot punish NYT from reporting "A man from NYC reports UFO sighted over Hudson River and it's getting viral". You cannot until you do editorial work.

This is what journalism is all about - not an algorithm's fault. Even NYT isn't immune from inaccurate reports. A responsible journalist has to do research, sometimes, days of research and dozen of interviews, to get a story out. Not simple "oh this is a fake news". No algorithm in the world can solve this problem. Never. Perhaps 1+1 = 3 is a fake news? No. Someone can write a joke about 1+1=3. In fact, we know humans invent the "1". We could have written 1 as "2". So 2+2 = 4 is no longer true. If we could use an algorithm to detect something is a fake news, an improved lie detector would have been produced. Remember history and scientific facts can be rewritten. What's is true today isn't necessarily true tomorrow.

https://en.wikipedia.org/wiki/New_York_Times_controversies

So why are they calling their product Google News? Why not Google Rumors, or Google Gossip? or "Stuff our Algorithms find to be popular or current but may or may not be true"?

Could you explain why it is good for society for Google to not give a shit about what it promotes?

I mean, the point of limited liability is not to give investors a path to profit regardless of all other things, it is to encourage investments that are beneficial to society. If it isn't working, we should retink it.

Because the money they make compared to how much content their serve is miniscule, and the only reason it works is scale. So if you want YouTube/Google/Twitter/Facebook to exist, you have to accept best effort algorithmic solutions, otherwise these services are not commercially viable and will cease to exist.

Can we ask these companies to do better? Sure. Can we expect them to be perfect? Not really.

If they are harming society they don't need to exist.

Google could trivially spend $200 million a month on moderation and other editorial functions, they just would rather not.

Net-net, I think Google's benefits outweigh it's issues. Economic analysis basically backs that up.

I admit I haven't run the numbers on Google, but when I looked at YouTube, requiring human minimum wage moderators to review every video would cost billions of dollars a year - more than the entire site's revenue.

So, sure, Google could drop another 200m on moderation, but I don't think that would get you to full moderation, so you still need algorithms in there somewhere.

If you want a moderated service use one. Or build one.

If Google becomes one (which they appear to be doing in some ways) I'll look for something that is more of a raw aggregator.

Google doesn't need to be a social engineer. Nor do many of the others who have decided they know what is "harming society and doesn't need to exist". Like a F*n small town preacher on a crusade to remove a book with dirty words from the library that is.

Indeed, restraining mega-corporations and burning books are exactly the same kinds of thing.

What if the books are made by mega-corporations "just out to make money"?

Attempting to restrict what can be published/viewed by do-gooders who think they know best is hardly a new gig. It just has different flavor now.

Who would moderate the moderators though? Didn't Facebook do this and had a problem with left-wing bias in the news they served?

> I mean, the point of limited liability is not to give investors a path to profit regardless of all other things, it is to encourage investments that are beneficial to society. If it isn't working, we should retink it.

From my understanding, a limited liability company exists to protect the business members and shareholders from legal action of their private persons. I.e. if the business goes under due to a poor product or something, they can't be held personally liable. That doesn't mean they are excused from criminal behavior, they can still be liable for some of that.

However, in either case there is no "benefit to society" clauses. It's simply a legal construct to protect people. In this case, Google did what they did... They probably weren't criminals though

Special legal protections that make corporations more efficient aren't natural rights, if they don't come with an overall benefit to society there is no reason to tolerate them.

I think they probably do come with an overall benefit to society, but the idea that giant corporations are going to do what they are going to do is one worth fighting against.

Yes, exactly. The point of Limited Liability is not simply to create a guaranteed path to profit (in fact it does no such thing; LLCs go bankrupt every day, and members can lose every penny they have personally invested).

Though there is no explicit "benefit to society" clause, that is the basic rationale for limited liability. It's an incentive to allow members to define the extent of their personal financial exposure. In so doing, they are more willing to invest in businesses, which can produce a net benefit to society (e.g. by employing people, paying taxes, etc.).

The marketplace of ideas can't work if we let a single entity decide which ideas are or aren't allowed to propagate. We all end up much better off if people are free to spread what those in power consider falsehoods; dealing with the occasional outright falsehood is the price we have to pay.

Google is already deciding which ideas propagate.

It's entirely reasonable for us to ask them not to do a shit job of it.

> Could you explain why it is good for society for Comcast to not give a shit about the content of the data it transmits?

It's not really a good comparison. Google purports to promote topical news in several places (search results, news., the Google App).

If Comcast was taking up an editorial function I'd criticize them for doing it poorly too.

Comcast says they provide you with "the internet". Google says that they're providing you with the news. News is understood to be a factual piece of information by any common measure of understanding, as opposed to rumors and gossip.

yes but would not AI be an interesting thing to apply to something like this? So the question is why Google is not yet investigating such an application of AI...its not going away..China has started using some of the same techniques as Russia

>Google will try to police its results better after this incident, but they can't effectively do this without mass censorship.

It's not mass censorship. It's editorialization. Something we've lost an appreciation for in our fetishization of a dystopian future that's always just over the horizon.

Facebook tried doing this to combat actual fake news (the one that ironically comes from the political segment in the U.S. most likely to cry wolf about it). This process was then characterized as facebook policing political content in a biased manner -- despite the fact that blatantly false content happened to continually come from sources with a conservative bent more often than not, which isn't Facebook's problem to deal with.

Point being, it's not censorship. And I wish any attempt to fight false information wasn't disingenuously painted as these organizations having some kind of political agenda or as "censorship". We need content aggregators of every kind (facebook, google, reddit, and even HN) to be more editorial, especially and specifically when content is presented in a manner that might help how credible that content looks. This whole wish-washy "oh just let the people decide what they wanna see, it'll work out in the end" with a nice touch of "all opinions matter" clearly hasn't and isn't working.

> Facebook tried doing this to combat actual fake news (the one that ironically comes from the political segment in the U.S. most likely to cry wolf about it). This process was then characterized as facebook policing political content in a biased manner -- despite the fact that blatantly false content happened to continually come from sources with a conservative bent more often than not, which isn't Facebook's problem to deal with.

Assuming we followed the same Facebook story (that intentionally set up boards of people - that were biased - to curtail the news in various ways), it seems like what you are trying to say is that alternative and conservative news is "fakes news", and they should receive algorithmic censorship.

Personally, for myself, the last 10 stories I've seen on "neutral" outlets like CNN and to the left outlets like Washington Post / New York / MSNBC - have been 100 percent driven by agenda rather than facts.

And the only time I'm able to get anything real is when I go to the other side, in which case all I have to do is tone it down 50%, yet on the left I have to literally invert what they are saying to get at the truth.

So how would you categorize which one of us is the idiot whose sources of news should be curtailed? Because I'm actually fine with you listening to CNN.

You're outnumbered on HN.

Some people are down voting him for name calling. Both republicans and democrats suffer from a persecution complex on HN. It's kind of funny.

All sites are agenda driven. But somehow sites with agendas that I disagree with are more guilty of bias. Funny how that works. Fake news, on the other hand, is claiming global warming is hoax invented by the Chinese. I would say only 10% of my facebook friends are conservative and I've only seen them post blatant lies like photoshops of black NFL players cheering the burning of a US flag. The liberals, at worst, get a little too excited about the Russian investigation. Though that has as much merit as Hilary's emails.

When someone says that I cannot discern fake from real, and for my own well being I should be protected from "fake news" (otherwise I'll start to believe that global worming is real / not-real), than I'm being treated like I'm some type of an "idiot", so why hide that word.

> Some people are down voting him for name calling.

I was surprised by the downvotes to a fairly argued comment, and then I read this hours ago and it seemed valid. The most upvoted comment in response now, however, calls him 'idiotic' several times throughout.

You're right he is, but if being outnumbered opinion-wise now leads to downvotes, I feel skeptical of HN's future.

Would we still upvote those who deny climate change for discussion's sake, for example?

> Would we still upvote those who deny climate change for discussion's sake, for example?

Why would you do that? There are plenty of opinions that are primarily "discussed" in bad faith, and that's one of the more significant ones.

Historically, if you look at past discussions, we've had great conversations by upvoting any comment that adds a new perspective to the mix.

If an ideal society wishes to eliminate climate change denial, for example, it won't be through silencing and stigmatizing their beliefs; it'll be through respectfully listening and debating.

As an example, I was rather unaware of the debate against climate change until reading a discussion on HN from people who were oft characterized as 'deniers.' One, for example, believed there were more immediate problems to be solved, and the other wasn't convinced by the science. It was a productive discussion, and I feel that everyone is better off having it, yet I fear that it wouldn't be condoned today without a mass of downvotes.

At this stage of the matter, it is quite possible that you fell for exactly the sort of bad-faith 'discussions' that pjc50 mentioned. Two forms of this are the repetition of thoroughly-debunked arguments and the injection of non-sequiturs (as I don't know what you are specifically referring to, I cannot say for certain, but most of what appears in this domain falls into at least one of these categories.)

I agree that there is no point in downvoting this sort of thing, but it is unfortunate if it gets upvoted by people who are not in a position to judge its validity. Scoring arguments on style is not the way to go (even if that is how the SAT essay is scored.)

Reports that the Texas gunman was Muslim were fake news in the truest sense; no relation to the truth and used primarily to promote racial hatred.

Does anyone have any sources of major news outlets reporting that the Texas gunman was Muslim? The problem is treating tweets as "news" and this is a problem by media outlets left and right. There are all kinds of tweets that are false and misleading that trend in both right and left circles why do we only complain about one side.

The tweets exist on "both sides." Sure. But how effective and endemic are these misleading tweets in each sphere?

I'm having a real hard time thinking of analogous lies on the left that have readily observable support or momentum in comparison to misinformation like the origin, religion, or race of a mass murder and/or claiming that an elected president wasn't actually born in the U.S.

Hard to dismiss such misinformation that's heartily consumed and propagated in whatever sphere as just being "tweets" or saying "it exists over there too!" when the effects of said misinformation in each sphere are significantly different.

Major news outlets, no. Not so major outlets, sure.

All the Googling I've seen seems to point to a "Santa Monica Observer" as the primary source of this claim. (See the Snopes article: https://www.snopes.com/texas-samir-al-hajeed-sam-hyde/ ), Snopes also links to a "Freedom Daily" and there are a couple of other sketchy news websites with this claim.

The Snopes article also mentions that a comedian known for trolling-style pranks (Sam Hyde -- https://en.wikipedia.org/wiki/Sam_Hyde#Hoaxes_and_pranks ) is tangent to these claims. So my best guess is that you are correct, someone treated a trollish tweet or post as actual news.

I'm not sure why bias is brought up all the time in "fake news", because bias is not the same as fake. Nobody should have issues with conservative news that is known for better vetting, even if their bias is strong (your Wall Street Journals, National Review, etc.). Fake is fake whether it comes from the right or the left. It is true that conservative fake news was much more prevalent in the 2016 election (https://web.stanford.edu/~gentzkow/research/fakenews.pdf), but "progressive" fake news definitely exists.

>it seems like what you are trying to say is that alternative and conservative news is "fakes news", and they should receive algorithmic censorship.

That's an incredibly disingenuous if not an outright idiotic comprehension on what was stated. If a source continually propagates blatantly false news it doesn't matter what the political bent is. It's going to lose credibility. Just like if someone is continually caught in a lie, you might stop listening to them, regardless of what race, religion, sex, or gender they are. You're not censoring them, you just know taking them at their word is a losing gamble.

The fact that a host of popular conservative sources continually participated in spreading false and misleading information isn't the fault of any content aggregator, and in this case Facebook certainly wasn't at fault. In fact, it seems like a problem that should be earnestly addressed within the conservative sphere, instead of conveniently using it to bolster a political platform and help with general ad revenue.

If you want to spit out little pathetic buzzwords and engage in intellectual lukewarmness, that's your prerogative. But you're going to have a hard time attempting to cry victim, or censorship, and engage in false equivocations when other people or organizations stop taking these sources seriously and choose to stop lending them a platform to spread lies and conspiracy theories.

> yet on the left I have to literally invert what they are saying to get at the truth.

Okay, well. I didn't realize I was dealing with unmitigated idiocy here. I'd have saved myself the key presses.

> outright idiotic comprehension

> spit out little pathetic buzzwords

> dealing with unmitigated idiocy here

With all due respect, your post contains some content, but this isn't behavior for HN and I'm really surprised it's being tolerated, let alone upvoted. Do you somehow expect to change any opinions with an insult-riddled comment like that?

So two things. You quoted three things, all 3 five words or less. At most that leaves 15 words across 3 paragraphs. Hardly worth drawing attention to.

Not that it really matters, because everything that was stated was pretty justified. It's curious that you take issue with my post or its tone, and not that out right dishonest response I was negating.

Looking at your posting history, and your continually false and misleading statements regarding all things geopolitical, I can only assume you're here to cause a stir. With all due respect.

I agree with your content but completely disagree with your delivery.

Did you come here from reddit? It's unnecessary and inappropriate, and I don't think that kind of behavior is 'justified' in response to someone you consider uneducated.

As for my (respectful) posting history: I only post when I feel that something has yet to be pointed out by anyone else, which will naturally be the more uncommon opinions and/or devil's advocate type of questions.

>Did you come here from reddit?

This is directly against the rules.

>It's unnecessary and inappropriate

No it's pretty appropriate. I didn't say anything uncalled for. Especially when someone is being dishonest. Congratulations on not being able to discuss the subject matter in your attempt to completely derail the conversation.

>devil's advocate type of questions.

Or just misinformation.

A fake news site owner said conservatives are more likely than liberals to believe and spread fake news. In fact, he said fake news didn't take off with liberals at all. I want to verify this. How easy would it be too replicate this experiment?

Edit: Found some links. Couldn't find the original article though.

0. Republicans seem more susceptible to fake news than Democrats (but liberals, don’t feel too comfy yet)

http://www.niemanlab.org/2017/05/republicans-seem-more-susce...

1. Article that says both left and right visits his fake news site.

http://www.politifact.com/punditfact/article/2017/may/31/If-...

2. Democrats are falling for fake news about Russia

https://www.vox.com/world/2017/5/19/15561842/trump-russia-lo...

3. The rise of left-wing, anti-Trump fake news

http://www.bbc.com/news/blogs-trending-39592010

I think that is going to be really difficult to discern with any accuracy. The different mindset of the two parties will likely lead to differences in the types of things that each will believe, but I don't believe that either group is immune to believing made up stories.

I have also found that the losing party tends to suffer from the wildest delusions. I'm not really sure why that is, but it is always sad to watch.

>I have also found that the losing party tends to suffer from the wildest delusions.

I would disagree with you on this statement. The far right is consumed by conspiracy theories (Seth Rich, Pizzagate etc) despite controlling all three branches of the Federal Govt.

Very hard to verify, but you could probably dig up some related data if you tried hard enough. Possibly things like fraud reports?

I was thinking of making my own sites to spread fake left and fake right news and tracking clicks.

> A fake news site owner said conservatives are more likely than liberals to believe and spread fake news. In fact...

Why would you believe what a "fake news" site operator tells you?

Smug "I am too smart to be fooled" liberalism, I'd guess, based on the fact this legitimate point is currently being downvoted.

I don't. It matches anecdotes at best. That's why I want to verify it.

That makes sense.

Ironically though, just as another anecdote (and to the best of my knowledge), the very first "fake news" right-wing site who's operator was confronted, ended up being a left-voting liberal who [paraphrasing] "hated conservatives and people on the right, and wanted to spread disinformation to them to pollute their arguments".

But I guess we'll never know, so the best solution is still non-censorship.

It's being down voted because it's an accusatory, non-answer to a speculative statement. The parent comment literally stated "I want to verify this." It's not a legitimate point in any sense of the word legitimate.

I'd have myself a hard think about what lead you to writing such a comment up.

Its not planned, it just works out that way.

I find it appalling that you don't even confront the question of "why are we using algorithms to promote news"?

Unless Google can somehow find a way to reliably present factual content, what they're doing here is fundamentally irresponsible.

That their choice of editorialization strategy is based on algorithms is entirely besides the point. The proof is in the pudding, and Google's pudding sometimes has chunks of shit in it. The onus is not on me to understand why that shit got there to be able to state that I don't want shit in my pudding.

But isn't the problem that Google isn't a news outlet? It's essentially showing us all the 'shit' that's floating around on the internet. Going to them expecting to get 'just the facts' is a category error.

Except that's exactly the expectation people have of news, and Google calls what they do news. It absolutely isn't a category error on the part of consumers when Google puts themselves in the category.

That's fair, to a point. Obviously most of what Google does is search, but "Top Stories" is cheerfully vague. It isn't Google News proper, but I agree it's intentionally confusing.

>I find it appalling that you don't even confront the question of "why are we using algorithms to promote news"?

So you think news could be effectively be curated on Google's scale by humans alone?

>Unless Google can somehow find a way to reliably present factual content, what they're doing here is fundamentally irresponsible.

Where do they claim that the content they present if factual? You are assuming this. They merely aggregate data, what you do with the data is your responsibility, not Google's. A higher ranking in results does not mean a higher level of accuracy in the content of the result, as anyone researching anything could tell you.

>The proof is in the pudding, and Google's pudding sometimes has chunks of shit in it. The onus is not on me to understand why that shit got there to be able to state that I don't want shit in my pudding.

I think what you are trying to say here is that you search for things expecting "the truth" to be given to you. Google only gives you things to read, you actually need to read them to determine if you think they're true or not. The onus is on you.

Don't believe everything that Google "tells" you. For a while, asking Google Now how tall the current president was (after Trump took office) would still give you Obama's height. Had you not actually looked at your phone, you wouldn't have noticed that the height it spoke back to you was for the wrong president. If you went around claiming Trump was the same height as Obama (they are actually pretty close, but that's besides the point) then you would be wrong and the responsibility would be solely yours for simply accepting whatever Google told you without checking other sources.

News is understood by the average person as a factual piece of information as opposed to a gossip or a rumor. They call it "Google News" and not "Stuff our algorithms find to be popular or current which may or may not be true!".

As a near monopoly, Google needs to do better than just defer to algorithm. When I searched for the Texas gunman name the first two articles on amp had headlines that he was a Muslim and Antifa. This is literally fake news and Google should be able to prevent gaming of it's algorithm from 4chan types

> and people who don't understand why Google employees did not manually review just one out of one hundred thousand things that the search engine indexed that day are suprised.

This drumbeat in recent news where one Silicon Valley firm or another gets caught presenting content designed to illicit response either through outrage, sensationalism, confirmation bias, or whatever via algorithms designed by behavior scientists that find offensive, misleading, false or outright maliciously incorrect content for that purpose which is done via an automated process and then posted to the front pages of various places where people are known to seek information after which the firm apologizes profusely, reminds people that this is just computers finding stuff, and tells everyone not to worry...

...is getting really, really fucking old. You know damn well what your algorithm is looking for. You know damn well it's good at finding it. You know damn well what it's going to do when it finds it. The fact that it's done automated-ly is not an excuse to LET IT KEEP FUCKING DOING THAT and just throw up your hands as if it's something not intended and not designed to do exactly that while you continue raking in the ad revenue because of it.

Google's the one in this story but it's far from the worst or only offender. How much longer exactly are we thinking on allowing Silicon Valley firms that make their money on promoting sensationalist garbage to continue to throw up their hands like it isn't exactly what they were meaning to do in pursuit of "eating the world" and walk away without so much as questioning if that may not be the best course of action?

>Of course, Google will try to police its results better after this incident, but they can't effectively do this without mass censorship. If you don't believe me, see the recent YouTube drama over monetization.

Well, if you call your product "News" and you promote fake news, that's pretty much a huge sign that you're not ready to ship yet.

I feel like there's a big push to blame "fake news" and related phenomena on tech companies right now.

Fake news is created and shared by people. Tech is just one of the vehicles through which we share it.

The problem isn't tech, it's people. Fix the root cause with education and the tech will reflect it.

"The problem isn't tech, it's people"

This sounds remarkably like an argument I've heard before, but there will always be destructive people and people seeking to profit from events like this.

In this case certain features created by tech companies perpetuate and amplify the destructiveness of those people, and that's a problem tech needs to solve quickly. Education could definitely help people filter information better, and in turn tech companies can leverage that curation to deliver better products.

The short term fix? Perhaps Google should just disable their tweet carousel (at least during certain events) until it works reliably and doesn't contribute to misinformation on this scale.

I would say that certain technologies perpetuate and amplify the voice of people, but not necessarily their destructiveness. Correct information is equally perpetuated and amplified. The ratio of correct to incorrect information seems unaffected. For this reason I am not convinced there's an actual problem.

I think the problem here is that Google's tweet results actually have been shown to display a fairly high ratio of incorrect information at some particularly critical times. I don't know why you'd say "the ratio of correct to incorrect information seems unaffected" - clearly the number of people exposed to incorrect information would be lower if Google did not show those tweets?

That's not surprising since they're essentially scraping Twitter. I'm not arguing that people shouldn't have a voice, but perhaps leave it to Twitter to rank tweets until Google's SNR ratio is better. Twitter just have a lot more signals to rank their users' tweets by (and for instance better assess whether tweets are being promoted by a botnet/voting ring).

Does Google not also display counter claims from tweets in its results?

If it can't discern fact from fake, it's hard to see how it wouldn't.

Do you remember the bad old days of website SEO, way back at the dawn of time? Back then someone could have just asserted that "good" sites will obviously appear as often as "bad" ones in results, if the search engine is unable to tell them apart. But of course that presumes either the "bad" sites aren't doing anything to game the search rankings, or that the "good" sites are just as effective at gaming the rankings, and we know for a fact that neither of these was the case. It took serious effort on the part of search engines to weed out the "bad" results, promote the "good" ones and close off the most effective ways to game the system.

We are going through that all over again now. Social media -- and yes, this includes important parts of Google search now, since they've decided to include social-media platforms prominently in results -- can be gamed the same way SERPs got gamed. And instead of just being gamed by people who want to sell you things (as in the bad old days of SEO), now they're also being gamed by political entities for the express purpose of swaying the course of nations.

These platforms wield enormous influence; it's not out of the question that, say, Twitter could literally be in a position to prevent or allow a major war. But the people who own and run these platforms are absolutely convinced of their own brilliance, and absolutely convinced that they could only be beaten by someone who has better tech skills than they do, when they're currently being played like a fiddle by people who have old-fashioned media and PR skills.

One of the fundamental social problems with Twitter is that counterclaims and retractions don't generally go viral in the same way that bullshit does. It's probably not one that you'll hear journalists talking about much, because it would make their own screw-ups look a lot more harmful and destructive.

it sounds sort of like guns don't kill people, people do. doesn't it.

Does that make it false ?

It certainly makes it suspect by being so clearly related to a phrase created to distract and confuse. (Guns kill people, many people, in accidents.)

I don't find it distracting or confusing. It's a rather simple statement. If you are worried over accidental deaths then we should consider discussing what we're going to do about the car accident problem before gun accidents.

If you are concerned about death then obesity kills 100x more people than guns.

And abortion kills 2x as many as obesity.

I mean, I understand this is your interpretation, but I feel that you exemplifying how the statement completely derails a discussion just proves my point.

Unfortunately not a brilliant analogy until guns start algorithmically deciding who to shoot at.

> Tech is just one of the vehicles through which we share it.

It's the fastest, most powerful, and most virulent vehicle we've found for it yet. This seems to be worthy of independent consideration, as we may have unknowingly crossed some sort of threshold.

As powerful as it is for sharing misinformation, it is equally powerful for sharing correct information.

When involved in a debate, if someone says the wrong thing, the audience does not interrupt and say "Moderator, you let the wrong thing be said. The debate must cease and you are fired." The correct thing is simply allowed to be said afterward.

There is a subtle draconian undertone to this anti-fake news rhetoric. We want tech to be responsible for our own actions, to such an extent that we are willing to sacrifice the natural flow of point/cointer-point of human discussion. Something about it just smells.

It’s not as powerful for sharing correct information.

There are plenty of credible studies showing how people have a bias to believe things which confirm their beliefs.

That means it takes more information to convince someone something they believe is wrong.

So base fears of “the other” are always going to be easy targets for fake news purveyors.

Right, but this isn't about point/counter-point or any kind of civil debate, it's about straight-up constructed lies and actual information.

Also, debate moderators have increasingly been calling out falsehoods during debates. There's a difference between being "wrong" and intentionally lying. I think it's great if the audience and moderators shut down verifiable lies, but of course wouldn't want them to shut down people I just disagree with.

The "algorithm" takes that kind of meaningful exchange and extracts the parts that are controversial (as judged by other users behavior). Those parts of the exchange is what will be presented to you, in hopes of an emotional response.

In the end, I suppose the human condition is to blame. After all, the "algorithm" is based on the collective actions of its users. But the tech companies facilitating this race to the bottom share some responsibility as well. Profiting (intentionally or otherwise) on the darker aspects of human psychology isn't unproblematic.

The problem is that people trust Google. If we go down the path you suggest, it means no one should ever trust what they find on Google. That throws it back 10 years to where it was just a collection of links and would potentially harm Google more than enforcing higher standards.

I trust google to return things that are relevant to my searches. Whether those things are true or false is up to me to determine after clickthrough.

This expectation that search results equal truth seems like a very recent assumption closely linked to this fake news development. I've never personally had that assumption myself and I am surprised to suddenly hear that it is apparently so common.

But, that's correct, people shouldn't trust what they find on Google. There are plenty of topics you can search in Google and get mostly false claims in the top results.

Even if Google was perfect, the problem is that behavior is desirable sometimes for users. Most people searching about bigfoot probably want a bunch of bigfoot sightings or conspiracy theory stuff, not "no, there probably isn't a bigfoot".

> The problem isn't tech, it's people.

Can you imagine if, circa 1900, we had taken this attitude toward workers getting injured in factories? Blame the worker for allowing their arm to be crushed, rather than the machine for having no safe guards? If we'd gone that route, then industrial accidents would still be one of the leading causes of death, rather than the minor cause of death that it is now.

Of course, some business leaders did blame the workers, but leading companies, prodded by the government, looked to make safer procedures as well as safer machines. And so the accident rate came down dramatically.

What's happening right now with the tech companies is similar, though it is social in nature. Over the last 20 years people have allowed tech to play a central role in their lives, and they've learned this can have costs and benefits, and now they are looking for ways to reduce those costs. And the costs are social in nature, and the tech can be altered to reduce those costs.

Although I agree with the solution of creating safeguards for the machine, I'm sure the accident rate would've also went down over time as people learned how to avoid them (though not as drastically).

There's a huge difference between making a machine safer and filtering news, however. The only downside to safeguarding the machine is the economic cost. With filtering news, it could end up rather dystopian and filter to fit a more complex agenda.

We're very cautious in America of any power we designate to a company or especially the government, as every power we assign is only beneficial situationally. When the situation changes, that power will likely stay with that entity, however. Our avoidance of giving away power could be classified as a bend-but-don't-break strategy, in regards to everything from news censorship to constitutional limitations (i.e. gun control).

The problem is that tech companies aren't just showing you everything everyone says... they're picking and choosing what to show you.

The 10,000ft overview is that they show you whatever keeps you coming back or staying on-site longer. This gives them more ad impressions and makes them more money.

Therefore, if their algorithms determine that showing you fake news, lies, and other socially-destructive and false propaganda is more profitable... they can and will do that.

So the question is, how do we regulate what the tech-utilities get to filter from your view? Should we regulate? It's not the regulation of speech, but rather the regulation of the willful filtration of speech for profit.

i.e., "dont show this guy dissenting views to his position or he'll stop coming back", where does that fit into our needs as a society and how should our laws tackle it, if at all?

I just wish people were able to collectively give up social media. I completely abhor the stranglehold its placed on culture and media, especially for the younger generation. It has dumbed everything down into what will get clicks, and is extremely invasive to boot. The only thing we "gain" from having social media as a society is advertisement.... which is essentially a way to get us to consume goods we wouldn't be consuming otherwise (i.e. waste). I don't see anything good coming from replacing the more or less natural human social structure that evolved over millions of years increasingly with an app designed to get us to buy goods we don't want, or get just 5 more minutes of attention. Such a waste of humanity's minds and effort.

Ultimately I really don't think you can say that Google/FB's dominance is a net positive for humanity. Sure we get to enjoy a lot of "free" goods, but those goods are merely a vector through which advertisements can learn more about us. Google/FB wouldn't be rolling in it if advertisers weren't making money off from advertising through them.

I see it almost as a maladaptive coping mechanism adopted by our society at large. We have a completely self-destructive reliance on perpetual inflation-beating growth, and at this point some of the most important mechanisms driving this growth are highly sophisticated advertisements designed to squeeze every last bit of consumer spending out of the economy as is possible. Not only that, but they've collectively convinced almost all of humanity to whore out all of their own personal information just so they can browse meme pages or check up on old high school acquaintances. It's absolutely disgusting behavior that to me resembles the informed preying on the ignorant.

A couple points and devil's advocacy on a couple as well:

Search is as bad as social, even if we give up social, it'll be hard to replace search (I read Wikimedia is working on something... a nonprofit search would be a fantastic asset to the world.)

Advertising generates a lot more than waste, in fact it may advertise to you a tool that you didn't know existed that prevents waste. Maybe rechargeable batteries or a water filter so you don't buy water bottles or who knows what, that sword cuts both ways. The ethics of advertisers and the lack of regulation in advertising though is easy to attack. That stuff is a mess.

As for Google being a net negative, I lean to your side of the argument but I think it'd be difficult and both sides would have a lot of merit; other search engines are just less efficient; think of all the man-hours of wasted research Google has saved by giving you the right result the first time, it must be billions of hours annually, maybe more. A lot of time and money has been spent competing, and the competition hasn't gotten much closer in terms of actual search results, so, if Google didn't exist, we may just be at a loss for that stuff.

Neither is intrinsically bad for society, merely the monetisation model. "State owned" search/social would obviously be even worse, however, so there isn't a clear path to having a version of either of those things which actually have as their primary aim the things Google/Facebook claim to.

> I just wish people were able to collectively give up social media.

As long as you realize that Hacker News is also social media.

I guess it's not social media itself, but its externalities and perverse incentives that are the problem. Hacker news isn't designed to be addictive nor do they sell my information to advertisers, so I don't have a problem with it

"I just wish people were able to collectively give up social media."

And video games and TV and radio and phone and books and writing. I do think the most interesting and satisfying way to do things are in person. Human physical and cultural evolution is having a hard time keeping up with technological changes.

"May you live in interesting times." Ancient Chinese curse.

> they're picking and choosing what to show you

Regulation is easy. When you wield editorial control over what people see, you should also gain some liability for any problems your decisions cause. If you don't to have that responsibility, then you need to act like a common carrier.

> Regulation is easy. When you wield editorial control over what people see, you should also gain some liability for any problems your decisions cause. If you don't to have that responsibility, then you need to act like a common carrier.

So you think that Google should not be attempting to filter out fake news at all? Or just, should not be ranking search results at all? I don't know how else to interpret "act like a common carrier", but I don't see how either of those would be beneficial... to anyone.

They can:

Decline to host any content from governmental agencies. Ie ban propaganda.

Instill some kind of fact checking. Maybe only post news from whitelisted domains. Make a digital equivalent of a press pass.

This could work for news.google.com but not for generic search. Unless google goes to a very restricted white listing for search. That would be "evil" in my book and I would try to stop using them for anything.

Interesting ideas about regulation from The Regulatory Review: https://www.theregreview.org/2017/10/10/cramer-bias-computer...

I don't blame Google for fake news. However, modern mass media amplifies fake information in near real-time, so it would make sense to also limit fake news from spreading so rapidly and possibly trigger consequences from those whose emotional capacity clearly exceeds their rational capacity. In other words, idiots going vigilante before the facts are known.

In one way, it's a good thing news can come from anywhere and reporting is more "democratic", but clearly we also need to find solutions to filter misinformation while at the same time safeguarding freedom of information in general. And that means cooperation of those companies that enable (any) news to spread without reality checks.

Note that it is also in (particulary) Google's best interest to provide facts and not to rapidly propagate misinformation. Who is interested in using a search engine that provides false or fake information?

People who can reliably be convinced to buy advertisers' goods when presented with fake information in response to a search. 'Cure cancer with one easy pill' and that sort of thing.

Fake news will always be profitable to publish because it thrives on exploiting our biases for profit. You can't just fix "people," there's something in English we call the "human condition" and people have been writing about it for thousands and thousands of years. It's unavoidable.

Rather than relying on Average Joe to simply not spread fake news, we need to shame these companies for actively profiting off of spreading fake information solely because of its huge impact on people on a psychological level. You wouldn't be ok with your local newspaper publishing an entire page on why "black on black crime is the only reason people advocate for gun control" so why are you ok with a global publisher publishing this information to millions of people across the planet?

News brands are in a losing battle to give people a reason to go to their web site first before facebook, google, twitter, etc. They are practically obligated to take every chance they can to persuade people. Failing that, getting some news brands banned will at least reduce competition and free up some eyeballs.

What do you do when education is confirming the fake news?

This is happening now in everything from misinformation on the causes of the civil war, to global warming misinformation, to evolution etc etc.

What’s more, the people promoting these misinformation campaigns are getting better organised and more effective at it.

It's 100% a tech problem. Optimizing for clicks pushes users into echo chambers.

Try watching a few breitbart or fox news videos on youtube in incognito mode and see what happens to the main feed. You'll end up with recommendations for info wars, roger stone and a bunch of conspiracy theories. You'll even start getting targeted ads for fake educational content like PragerU (https://www.youtube.com/watch?v=h1jdBmQcwnw)

Facebook / Twitter / Google are happy to keep feeding you this crap because it leads to more clicks, engagement and ad views. With recommenders and social graphs shaping most of what we see on the web a lot of less tech savvy people end up being exposed to a more radical looking side of the internet.

A surprising amount of the actual viral bullshit I've seen out there, especially on Twitter, has been started and spread by journalists working for respectable publications. (Just recently they got a bunch of viral attention by misdescribing what Trump did at a public press op that was videoed, for example.)

But they claim to use their tech to do things like deliver news, ban fake news, control misinformation, etc. If they themselves are claiming this, do you want to hold them accountable?

Its the same thing as why people are upset with Fox News. ITs because they call themselves a news organization and are supposed to be fair and balanced. So obviously people are going to try to hold them to their own words. If they had just called their channel Fox's "made up crazy shit", I don't think there would be as big of a backlash against them.

Not sure what to think here... Google's 'public liaison for search' says "Google briefly carried tweets with dubious info "[0] without defining further what that means - except their Twitter results are changing "second by second"[1], and later assuring that it "only happened for a few thousands who searched for his name"[2].

This was not my experience. I took a screenshot of my search results at 9:06PM ET [3], e.g. 1-2+ hours after the egregious tweets (which were already prominently featured within 30-40 minutes, if not earlier, as several other sources suggest[4]).

I might be missing something, but it seems disingenuous to suggest that this only happened for a short period of time, while downplaying the matter by vaguely and generally hinting at second-by-second dynamic tweet update algorithms.

Also I even think the claim that only a few thousand people saw this may be disproven by Google's own statistics. According to Google Trends, searches for Devis Patrick Kelly peaked between 7-9PM ET [5] - so before I took the screenshot that night.

I find it very hard to believe that only a few thousand people saw this over the 2+ hours when this peaked on that night. Based on the Google Trends graph[6] it seems the vast majority of searches happened during the peak, and I suspect it's reasonable to assume that millions have searched for his name since his name was publicly revealed?

[0] https://twitter.com/dannysullivan/status/927713318172635137

[1] https://twitter.com/dannysullivan/status/927713426440253440

[2] https://twitter.com/dannysullivan/status/927713578827653120

[3] https://imgur.com/a/ao4kK

[4] https://www.theregister.co.uk/2017/11/06/google_twitter_fake...

[5] https://imgur.com/a/lAh12

[6] https://trends.google.com/trends/explore?date=2017-11-05T05%...

More than likely it is Google trying to downplay it, but it's their search being gamed - the BBC article is actually inaccurate as far as I'm aware, as at least in the case of Sam Hyde, I was under the impression he was in on the joke, though I may be wrong.

Regardless, the length of time probably isn't as important as the frequency with which this occurs, as the search frequently returning verifiably wrong or fake information is probably a bigger issue. The length of time it lasts is just a symptom of the tool providing news without the news being curated.

Yes, there is bad information on the Internet. Yes, when you aggregate tons of user content some inaccurate things will be there. Go somewhere else BBC, NYT, WaPo. We know, you'd like to see a content regulated Internet. I wouldn't and I'm weary of seeing these articles day in and day out.

Beside why are articles from newspaper on HN front page so much lately?

From HN guidelines: Off-Topic: "Most stories about politics, or crime, or sports, unless they're evidence of some interesting new phenomenon....... If they'd cover it on TV news, it's probably off-topic."

This self serving media policy seeking doesn't belong here IMOP. It's not interesting, it's not informative. Shouldn't be constantly on front page day after day.

>Google's Danny Sullivan - "We want to get this right"

I feel like in a better world that would be, "We are legally obligated to get this right or stop trying."

A little off-topic, but on similar lines: searching for "demagogue" on google, gets you a Trump's picture: https://i.imgur.com/qFJk9io.png

I wonder how that ends up happening, is google able to sum up popular opinion or is this someone's mischief?

I wonder when we finally get mandatory electronic locks on guns. Having an electronic license for a single gun could have prevented this man from using a whole battery of guns.

It makes little sense that we can have more sophisticated locks on an iPhoneX than on guns.

Reply on news.ycombinator.com